The invention is generally related to electronic sensing devices, and, more particularly, to sensors for sensing objects located near or about the sensor.
In the electronic sensing market, there are a wide variety of sensors for sensing objects at a given location. Such sensors are configured to sense electronic characteristics of an object in order to sense presence of an object near or about the sensor, physical characteristics of the object, shapes, textures on surfaces of an object, material composition, biological information, and other features and characteristics of an object being sensed.
Sensors may be configured to passively detect characteristics of an object, by measuring such as temperature, weight, or various emissions such as photonic, magnetic or atomic, of an object in close proximity or contact with the sensor, or other characteristic. An example of this is a non-contact infrared thermometer that detects the black body radiation spectra emitted from an object, from which its temperature can be computed.
Other sensors work by directly exciting an object with a stimulus such as voltage or current, then using the resultant signal to determine the physical or electrical characteristics of an object. An example of this is a fluid detector consisting of two terminals, one that excites the medium with a voltage source, while the second measures the current flow to determine the presence of a conductive fluid such as water.
Since a single point measurement of an object often does not provide enough information about an object, it is often advantageous to collect a 2-dimensional array of measurements. A two dimensional array of impedance may be created by moving a line sensing array over the surface of an object and then doing a line by line reconstruction of a two dimensional image like a fax machine does. An example of this is a swiped capacitive fingerprint sensor that measures differences in capacitance between fingerprint ridges and valleys as a finger is dragged across it. Such sensors reconstruct a two dimensional fingerprint image after the fact using individual line information.
A simpler way to obtain a two dimensional image is to create a two dimensional sensing array. Such sensors however can be prohibitive in cost due to the large number of sensing points needed in the array. An example of this is a two dimensional capacitive fingerprint sensor. A number of these are currently manufactured but use 150 mm2 or more of silicon area and are therefore cost prohibitive for many applications.
These different types of electronic sensors have been used in various applications, such as biometric sensors for measuring biological features and characteristics of people such as fingerprints, medical applications such as medical monitoring devices, fluid measuring monitors, and many other sensor applications. Typically, the sensing elements of the various devices are connected to a processor configured to process object information and to enable interpretations for object features and characteristics.
There are many applications for two dimensional image sensors as a particular example, and innovators have struggled with state of the art technology that has come short of desired features and functions. Fingerprint sensors, for example, have been in existence for many years and used in many environments to verify identification, to provide access to restricted areas and information, and many other uses. In this patent application, different types of fingerprint sensors will be highlighted as examples of sensor applications where the present invention is applicable for simplicity of explanations, but other types of applications are also relevant to this background discussion and will also be addressed by the detailed description of the invention. These placement sensors may be configured to sense objects placed near or about the sensor, such as a fingerprint placement sensor that is configured to capture a full image of a fingerprint from a user's finger and compare the captured image with a stored image for authentication. Alternatively, sensors may be configured to sense the dynamic movement of an object about the sensor, such as a fingerprint swipe sensor that captures partial images of a fingerprint, reconstructs the fingerprint image, and compares the captured image to a stored image for authentication.
In such applications, cost, though always a factor in commercial products, has not been so critical—accuracy and reliability have been and still remain paramount factors. Typically, the placement sensor, a two-dimensional grid of sensors that senses a fingerprint image from a user's fingerprint surface all at once, was the obvious choice, and its many designs have become standard in most applications. Once the fingerprint image is sensed and reproduced in a digital form in a device, it is compared against a prerecorded and stored image, and authentication is complete when there is a match between the captured fingerprint image and the stored image. In recent years, fingerprint sensors have been finding their way into portable devices such as laptop computers, hand held devices, cellular telephones, and other devices. Though accuracy and reliability are still important, cost of the system components is very important. The conventional placement sensors were and still are very expensive for one primary reason: they all used silicon sensor surfaces. These surfaces are very expensive, as the silicon material is as expensive as the material to make a computer chip. Computer chips, of course, have become smaller over the years to reduce their cost and improve their performance. The reason the fingerprint silicon could not be made smaller: they need to remain the size of the average fingerprint, and the requirement for full scanning of the users' fingerprints simply cannot be compromised. The full print is required for adequate security in authentication.
Enter the fingerprint swipe sensor into the market. Swipe sensors are fundamentally designed with a line sensor configured to sense fingerprint features as a user swipes their finger in a perpendicular direction with respect to the sensor line. The cost saver: swipe sensors need much less silicon, only enough to configure a line sensor with an array of pixel sensors. The width is still fixed based on the average fingerprint width, but the depth is substantially smaller compared to the placement sensor. Some swipe sensors are capacitive sensors, where capacitance of the fingerprint surface is measured and recorded line by line. Others send a small signal pulse burst into the surface of the fingerprint surface and measure response in a pickup line, again recording fingerprint features line by line. In either case, unlike the placement sensors, the full fingerprint image needs to be reconstructed after the user completes the swipe, and the individual lines are reassembled and rendered to produce a full fingerprint image. This image is compared with a fingerprint image stored in the laptop or other device, and a user will then be authenticated if there is an adequate match.
For the capacitive swipe sensors, the first generation sensors were constructed with direct current (DC) switched capacitor technology (for example U.S. Pat. No. 6,011,859). This approach required using two plates per pixel forming a capacitor between them, allowing the local presence of a finger ridge to change the value of that capacitor relative to air. These DC capacitive configurations took images from the fingerprint surface, and did not penetrate below the finger surface. Thus, they were easy to spoof, or fake a fingerprint with different deceptive techniques, and they also had poor performance when a user had dry fingers. RF (Radio Frequency) sensors were later introduced, because some were able to read past the surface and into inner layers of a user's finger to sense a fingerprint. Different radio frequencies have been utilized by various devices along with different forms of detection including amplitude modulation (AM) and, phase modulation (PM). There are also differing configurations of transmitters and receivers, one type (for example U.S. Pat. No. 5,963,679) uses a single transmitter ring and an array of multiple low quality receivers that are optimized for on chip sensing. In contrast another type (for example U.S. Pat. No. 7,099,496) uses a large array of RF transmitters with only one very high quality receiver in a comb like plate structure optimized for off chip sensing.
One key impediment to the development of low cost placement sensors has been the issue of pixel density, and the resultant requirement for a large number of interconnections between layers of the sensor device. A typical sensor for a fingerprint application will be on the order of 10 mm×10 mm, with a resolution of 500 dpi. Such a sensor array would be approximately 200 rows by 200 columns, meaning there would need to be 200 via connections between layers in the device. While semiconductor vias can be quite small, the cost for implementing a sensor in silicon has proven to be prohibitive, as mentioned above.
In order to produce a placement sensor at a low enough cost for mass market adoption, lower cost processes such as circuit board etching must be employed. The current state of the art in circuit board via pitch is on the order of 200 um, vs. the 50 um pitch of the sensor array itself. Additionally, the added process steps required to form vias between layers of a circuit board significantly increase the tolerances for the minimum pitch of traces on each of the layers. Single-sided circuits may be readily fabricated with high yield with line pitch as low as 35 um, whereas double sided circuits require a minimum line pitch on the order of 60 um or more, which is too coarse to implement a full 500 dpi sensor array. One further consideration is that at similar line densities, double-sided circuits with vias are several times more expensive per unit area than single sided, making high-density double sided circuits too expensive for low cost sensor applications.
For laptop devices, adoption of the swipe sensor was driven by cost. The swipe sensor was substantially less expensive compared to the placement sensors, and most manufacturers of laptops adopted them based solely on price. The cost savings is a result of using less silicon area. More recently a substitute for the silicon sensor arose, using plastic Kapton™ tape with etched sensing plates on it, connected to a separate processor chip (for example U.S. Pat. No. 7,099,496). This allowed the silicon portion of the sensor to be separated from the sensing elements and the silicon to follow Moore's law, shrinking to an optimal size, in length, width and depth in proportion to advances in process technology. Although this advance in the art enabled cheap durable Swipe Sensors, it did not overcome the basic image reconstruction and ergonomics issues resulting from changing from a simple two dimensional placement format. In addition to Swipe Sensors being cheaper, they take up less real estate in a host device, whether it is a laptop or a smaller device, such as a cellular phone or personal data device.
In most swipe class sensors, the fingerprint reconstruction process turned out to be a greater ergonomic challenge to users and more of a burden to quality control engineers than initially expected. Users needed to be trained to swipe their finger in a substantially straight and linear direction perpendicular to the sensor line as well as controlling contact pressure. Software training programs were written to help the user become more proficient, but different environmental factors and the inability of some to repeat the motion reliably gave Swipe Sensors a reputation for being difficult to use. Initial data from the field indicated that a large number of people were not regularly using the Swipe Sensors in the devices that they had purchased and opted back to using passwords. Quality control engineers who tried to achieve the optimum accuracy and performance in the matching process between the captured and reconstructed image found that the number of False Rejects (FRR), and False Acceptances (FAR), were much higher in Swipe Sensors than in placement sensors. Attempts to improve these reconstruction algorithms failed to produce equivalent statistical performance to placement sensors.
Other claims of the Swipe Sensor such as the use of less host real estate did not pan out. Various ramps, wells and finger guides had to be incorporated into the surfaces of the host devices to assist the user with finger placement and swiping. These structures ended up consuming significant space in addition to the actual sensor area. In the end, swipe sensors ended up taking up almost as much space as the placement sensors. This was not a big problem for full size laptops, but is currently a substantial problem for smaller laptops and netbooks, mobile phones, PDAs, and other small devices like key fobs.
Real estate issues have become even more of an issue with mobile device manufacturers who now require that the fingerprint sensor act also as a navigation device, like a mouse or touch-pad does in a laptop. The swipe sensor has proved to be a poor substitute for a mouse or touch pad due to the fact that they are constructed with an asymmetric array of pixels. Swipe sensors do a good job of detecting motion in the normal axis of the finger swipe but have difficulty accurately tracking sideways motion. Off axis angular movements are even more difficult to sense, and require significant processor resources to interpolate that movement with respect to the sensor line, and often have trouble resolving large angles. The byproduct of all this is a motion that is not fluid and difficult to use.
It is clear that low cost two dimensional fingerprint sensor arrays would serve a market need, but present art has not been able to fill that need. Conventional capacitive fingerprint sensors typically use distinct electrode structures to form the sensing pixels array. These electrode structures are typically square or circular and can be configured in a parallel plate configuration (for example U.S. Pat. Nos. 5,325,442 and 5,963,679) or a coplanar configuration (for example U.S. Pat. Nos. 6,011,859 and 7,099,496).
These prior art approaches cannot be configured into a low cost two dimensional array of sensing elements. Many capacitive fingerprint sensors (for example U.S. Pat. Nos. 5,963,679 and 6,011,859) have plate structures that must be connected to the drive and sense electronics with an interconnect density that is not practical for implementation other than using the fine line multilayer routing capabilities of silicon chips and therefore require lots of expensive silicon die are as stated before. Other sensors (for example U.S. Pat. No. 7,099,496) use off chip sensing elements on a cheap polymer film, but the sensor cell architecture is inherently one dimensional and cannot be expanded into a two dimensional matrix.
Another application for capacitive sensing arrays has been in the area of touch pads and touch screens. Because touchpad and touch screen devices consist of arrays of drive and sense traces and distinct sense electrodes, they are incapable of resolutions below a few hundred microns, making this technology unsuitable for detailed imaging applications. These devices are capable of detecting finger contact or proximity, but they provide neither the spatial resolution nor the gray-scale resolution within the body of the object being sensed necessary to detect fine features such as ridges or valleys.
Conventional art in the touchpad field utilizes a series of electrodes, either conductively (for example U.S. Pat. No. 5,495,077) or capacitively (for example US publication 2006/0097991). This series of electrodes are typically coupled to the drive and sense traces. In operation these devices produce a pixel that is significantly larger in scale than the interconnect traces themselves. The purpose is to generally sense presence and motion of an object to enable a user to navigate a cursor, to select an object on a screen, or to move a page illustrated on a screen. Thus, these devices operate at a low resolution when sensing adjacent objects.
Thus, there exists a need in the art for an improved device that can provide high quality and accurate placement sensors for use in different applications, such as fingerprint sensing and authentication for example, and that can also operate as a navigation device such as a mouse or touch pad. As will be seen, the invention provides such a device that addresses these and other sensor needs in an elegant manner.